Fungal parasitism of cereal aphids in South Africa.

Thesis (Ph.D.)-University of Natal, Pietermaritzburg, 2002.The Russian wheat aphid, Diuraphis noxia is one of the most destructive exotic invaders of South Africa, capable of reducing individual wheat plant yields by up to 90%. Entomopathogenic fungi are important natural mortality factors associated with this aphid in its Eurasian endemic habitats as well as in the United States and Canada; their impact often exceeding that of predators and parasitoids. The principal objectives of this study included (1) baseline characterization of the aphid-pathogenic flora associated with aphids from South Africa, with special reference to six common cereal aphids, (2) quantification of the comparative impact of the different fungal species on the cereal-aphid complex in three different wheat producing regions of South Africa, (3) field evaluation of the Hyphomycete Beauveria bassiana against D. noxia on resistant wheat, (4) screening of six fungicides for their potential usage in managing entomophthoralean epizootics within greenhouse rearings of the Russian wheat aphid, and (5) development and evaluation of a novel bioassay protocol for screening entomopathogenic Hyphomycetes against D. noxia. A total of nine species of fungi known to infect and kill aphid hosts were collected, including the six entomophthoraleans, Pandora neoaphidis, Conidiobolus thromboides, Conidiobolus obscurus, Entomophthora planchoniana, Conidiobolus coronatus and Neozygites fresenii, and three Hyphomycetes, Beauveria bassiana, Verticillium lecanii, and Paecilomyces farinosus. The former four entomophthoraleans are considered first reports from this country. For the first time, morphological characteristics of these nine South African-collected species are visually depicted and techniques for their isolation and in vitro culture discussed. Seven species of fungi were recorded from D. noxia, of which P. neoaphidis was the most important, causing up to 50% mortality during the late season under dryland conditions in the summer-rainfall region. Mycoses at epizootic levels, together with the large diversity of fungal species recorded from this host, indicated a high level of susceptibility to fungal infection. In contrast, infection of the oat aphid Rhopalosiphum padi remained < 5% despite favourable numbers of hosts and apparently suitable environmental conditions. This phenomenon strongly suggests some level of low susceptibility to fungal infection in this species. Under irrigated conditions m KwaZulu-Natal, the rose-gram aphid Metopolophium dirhodum was the predominant aphid but remained below economical injury levels. Field surveys revealed that this aphid was effectively targeted by P. neoaphidis and C. obscurus, and findings suggest that in some areas of South Africa entomophthoralean fungi effectively suppress this aphid, negating the need for insecticide applications. On average, ca. 61% control of D. noxia on resistant wheat was observed following an application of B. bassiana (5 x 10¹³ conidia per hectare) during the early flag-leaf stage of the wheat. Efficacy of B. bassiana applications on younger plants appeared to be influenced by the level of aphid activity, possibly explained by secondary pick-up of inoculum by D. noxia. These findings accentuate the importance of understanding the tritrophic relationship between host plant, pest and pathogen. The fungicides copper oxychloride, mancozeb + oxadixyl, captab + metalaxyl, bittertanol, iprodione, and mancozeb at a rate of 0.1% active ingredient moderately to strongly inhibited C. thromboides vegetative growth (mean inhibition 81.1 %). Mancozeb at concentrations of 10.0, 2.0, 1.25, 0.5, 0.08, and 0.016% was further evaluated in vitro. The fungus growth cut-off point, midway between 1.25 and 2.0% mancozeb, was calculated and a rate of 1.625% active ingredient per litre of water was used to decontaminate the fungus-infected D. noxia cultures. A novel bioassay protocol was developed, employing live host plants for rearing aphids post inoculation. Using this design against D. noxia, an average LC₅₀ estimate of 85 conidia per mm² for B. bassiana strain GHA was calculated. Control mortality was restricted to levels below 4%. The data indicated high precision due to an average coefficient of variation for slope of less than 20%, and an average chi-squared value of 5.46 ± 2.74 (n = 10 assays). The design will accommodate the use of cereal aphid species other than D. noxia, while live host plants will facilitate tritrophic studies on the effect of host-plant resistance on fungus-induced mortality of D. noxia